Growing Pleurococcus

The purpose of the growing Pleurococcus experiment is to find which side of the most common Pleurococcus tree, and why? I think that Pleurococcus is predicted to grow on the north side of the tree. I guess this happens because the southern part of the tree receives the most sunlight. The heat from this sun can dry Pleurococcus and it will dry out and kill. Because Pleurococcus is algae, we do photosynthesis. This means that it requires water as shown in the photosynthetic equation below.

In order to consider / generate the influence of "tree kind" "height" "light intensity" "preliminary method of observation of temperature" on the tree, investigate the growth factor of the trunk Pleurococcus, which side of the tree is north We use a compass to see if it is. "Then grind and grasp the body and find out how many squares.« Write the results on the table and repeat east, south, west Repeat Result Tree Number Site 1 North Site 1 East Site 1 South Site 1 West Site 2 North Site 2 East Site 2 South Site 2 West 1 25 24 4 25

So what happened in the background? To plant trees, you need to decide which functions to select, the conditions for segmentation, and when to stop. Since trees usually grow randomly, you need to arrange them so that they look beautiful. Let's start with the general technique used for splitting. Let's consider an early example of a tree learned from the Titanic dataset. For the first segmentation or route, all attributes / functions are taken into account and training data is grouped based on segmentation. Since there are three functions, there are three candidate splits. Now use this function to calculate the accuracy of the cost of each split. Please select the lowest cost segmentation by the passenger sex in our case. Since the resulting groups can be subdivided using the same strategy, the algorithm is inherently recursive. Because of this process, we have a desire to excessively reduce costs, so this algorithm is also called a greedy algorithm.

Most decision trees go through two phases: the tree growth (split) phase, and then the pruning phase. The tree growth phase is an iterative process involving gradually splitting the data into small subsets. Everyone is just considering the data in one node. In the first iteration, the root node containing all the data is taken into account. Subsequent iterations are applied to derived nodes that contain a subset of the data. An important function of the tree splitting algorithm is that it is greedy. Greedy algorithms make decisions locally, not globally. When determining the segmentation of a particular node, the greedy algorithm does not prefetch the tree to see if another decision yields a better overall result. In general, the initial division (divided closer to the root) is not the best segmentation compared to any local measurement, but it will result in a tree with higher overall accuracy when used.